Strategies for metabolite profiling based on liquid chromatography

Journal of Chromatography B, 1044 (2017) 103–111

Contents lists available at ScienceDirect

Journal of Chromatography B journal homepage: www.elsevier.com/locate/chromb

Review

Strategies for metabolite profiling based on liquid chromatography Javier Saurina ∗ , Sonia Sentellas Department of Analytical Chemistry, University of Barcelona, Martí i Franquès 1-11, 08028 Barcelona, Spain

a r t i c l e

i n f o

Article history: Received 22 June 2016 Received in revised form 17 December 2016 Accepted 8 January 2017 Available online 9 January 2017 Keywords: Metabolite profiling Structural elucidation Liquid chromatography Mass spectrometry

a b s t r a c t This paper aims at covering the principal strategies based on liquid chromatography (LC) for metabolite profiling in the field of drug discovery and development. The identification of metabolites generated in the organism is an important task during the early stages of preclinical research to define the most proper strategy for optimizing, adjusting metabolic clearance and minimizing bioactivation. An early assessment of the metabolite profile may be critical since metabolites can contribute to pharmacological and/or toxicological effects. The study of metabolites first involves their synthesis/generation and their further characterization and structural elucidation. For such a purpose, both in vitro and in vivo methods are commonly used for the generation of the corresponding metabolites. Next, analytical methods are used to tackle identification and characterization studies. Among the arsenal of techniques available in our labs, we will focus on LC, especially coupled to mass spectrometry (LC–MS), as one of the most powerful approaches for metabolite identification, characterization and quantification. Here, the topic of metabolite profiling based on LC will be addressed and representative examples of different possibilities will be discussed. © 2017 Elsevier B.V. All rights reserved.

Contents 1. 2. 3. 4. 5. 6. 7.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 Metabolite generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 Sample treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 Metabolite separation by liquid chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 4.1. Detectors in liquid chromatography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Detection, structural elucidation and quantification of metabolites by LC–MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 Chemometrics for the treatment of LC data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 Conclusions/further trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

1. Introduction The vast majority of drugs are transformed total or partially in the organism. Generally, such processes involve the formation of more polar species than the original drug to be more easily excreted. Besides, these processes commonly imply a loss of pharmacological activity together with a reduction of toxicity. Of course, exceptions to these general patterns sometimes occur in which some drugs may generate pharmacologically active or, even, toxic metabolites. For instance, enrofloxacin is a quinolone antibiotic

∗ Corresponding author. E-mail address: [email protected] (J. Saurina). http://dx.doi.org/10.1016/j.jchromb.2017.01.011 1570-0232/© 2017 Elsevier B.V. All rights reserved.

whose main metabolite ciprofloxacin maintains a high antibacterial effect [1]. The most exceptional cases of metabolic activation deal with the so-called prodrugs, i.e. originally inactive molecules that acquire their activity after undergoing metabolism. As a different point, the occurrence of toxicity induced by metabolites is another occasional (and unwanted) phenomenon associated to drug metabolism. As a result, dozens of potentially active molecules must be discarded during the early stages of research and development of new drug candidates due to such problems. Toxicity may ever be detected in approved drugs that are afterwards withdrawn from the market or subjected to restrictions because of unwanted effects often caused by reactive metabolites [2]. Most of the adverse effects linked to metabolites are dose-dependent and can be assessed from regulatory animal toxicity studies via

104

J. Saurina, S. Sentellas / J. Chromatogr. B 1044 (2017) 103–111

in vitro and in vivo assays. In a few number of cases, however, idiosyncratic adverse effects in humans may occur which are seldom predictable from assays with toxicological species [3]. In these circumstances, activity/toxicity effects of some drugs might differ significantly among patients due to inter-individual variabilities. New promising trends to deal with such differences rely on personalized treatments specifically designed for each patient. Among other possibilities, monitoring drug or metabolite levels in body fluids and designing controlled drug release systems may result in highly efficient strategies to minimize toxicities. The metabolic biotransformations can be classified into phase I and phase II reactions as follows. Phase I metabolism (or functionalization reactions) introduce functional groups in the structure of the drug. The most common phase I reactions consist of oxidations (aliphatic and aromatic hydroxylation, N-, O- and Sdealkyilation, oxidative deamination, N-oxidation, etc.), reductions (azo-reduction and nitro-reduction) and hydrolysis (of esters and amides). Phase II metabolism (or conjugative reactions) consist of the formation of covalent bonds between a reactive functional group of the raw drug or a phase I metabolite with endogenous compounds. Some important phase II reactions comprise glucuronidation, glycosylation, acylation, acetylation, sulfation, methylation and conjugation with amino acids, glutathione (GSH) and fatty acids. The identification and quantification of metabolites generated in the organism are an important issue in drug discovery and development. The preliminary evaluation of the metabolic profile is a necessary step to define the most proper strategy of drug optimization, adjusting the metabolic clearance and controlling the bio(de)activation. A comprehensive preclinical evaluation, including inter-species comparison may be important to assess the pharmacological and/or toxicological effects of metabolites. Anyway, it should be mentioned that metabolic profiles may differ notably depending on the species assayed. In this way, the variety of metabolites, the concentration levels and kinetics are often species-dependent [4]. In general, the incidence of metabolites on the drug activity or toxicity may be negligible if the metabolism rate is low. It is accepted that metabolites representing less than 10% of the parent systemic exposure in humans may be disregarded. Conversely, those metabolites occurring at concentrations above 10% have to be evaluated more thoroughly to assess their potential effects and risks. If human exposure to these metabolites is similar

or lower to that observed in toxicological animal species, no further preclinical tests are needed. However, special attention should be paid to those metabolites present at disproportionately lower levels in animal studies compared to humans [5]. A general scheme to be followed for metabolite profiling in the research and development of new drugs is shown in Fig. 1. The comprehensive evaluation of the drug metabolism comprises various main stages, namely: metabolite generation, isolation and purification, separation, identification and structural elucidation and quantification. According to Fig. 1, the study of drug metabolism first involves the (bio)synthesis or generation of metabolites for their subsequent characterization and structural elucidation. Preliminarily, in vitro assays with model systems of different complexity, such as microsomes, cells and tissues, can be conducted to obtain a first insight on the metabolite formation [6,7]. Electrochemical generation methods have recently been introduced to provide a straightforward and cheap picture of oxidative routes [8,9]. Anyway, in vivo assays are required for a more comprehensive and accurate evaluation of the overall drug metabolism. It should be mentioned that apart from experimental approaches, in silico methods based on prediction rules are increasingly used for a rough approximation to drug metabolism in order to quickly detect some potential sources of warnings [10,11]. Once metabolites have been generated, they need to be identified and quantified using the great arsenal of analytical techniques available in our laboratories. Often, preliminary sample treatments are applied to purify and preconcentrate the analytes based on protein removal, liquid and solid phase (micro)extraction, etc. The resulting extracts are then analyzed using powerful instrumental techniques such as liquid chromatography (LC) coupled to mass spectrometry (MS) and nuclear magnetic resonance (NMR) for performing separation, structural elucidation and quantification of the components of interest. Structural information of metabolites is essential in drug discovery to quickly identify and establish the metabolite profiles and the rate of drug biotransformation. However, metabolite quantification may result in a complex task as standards of most of metabolites are unavailable. This issue may be solved from the estimation of the drug decay or under the assumption of similar instrumental sensitivities for the parent drug and its metabolites. This is true in the case of radioactivity detection and is commonly accepted in the case of UV detection as the

Fig. 1. General strategy for metabolite profiling.

J. Saurina, S. Sentellas / J. Chromatogr. B 1044 (2017) 103–111

105

Fig. 2. Summary of methods used for metabolite generation.

absorbing groups of drug and metabolite molecules are often the same. In MS, anyway, the ionization efficiencies of these molecules may differ substantially so that more complex strategies are required for accurate metabolite determinations. This paper is an overview on the novel trends in metabolite profiling based on LC, with special attention in most recent advances on coupling with MS for identification, structural elucidation and quantification purposes. In order to focus the topic properly, sample treatment and detection possibilities are also commented in the following sections. Finally, the application of chemometric methods for the treatment of LC data is also introduced. 2. Metabolite generation The first step when dealing with metabolite profiling is the selection of the most suitable source of metabolites (Fig. 2). The so-called “in silico” approach can be used for a preliminary modeling of the principal metabolic biotranformations of a given drug [12,13]. Computational methods relying on bibliographic information and expert knowledge rules in metabolism are used to estimate the structures of the most probable metabolites as well as their relative abundances. In vitro assays comprise various experimental procedures using biologic material of different complexity, from enzymes to tissues

[6,7]. The simplest example relies on recombinant enzymes to carry out the drug incubation. Progressing on the complexity of in vitro assays, subcellular fractions including microsomes, S9 fractions and cytosol suspensions from liver, intestine or other tissues can be used. Among them, liver microsomes are the in vitro biological systems most commonly used. Sometimes, in order to increase the prevalence of metabolite generation, induced (eg. aroclor 1254) rat liver microsomes can be used. Liver microsomes contain the cytochrome P450 which is the family of enzymes responsible for about the 75% of the total number of different metabolic reactions. Other important metabolizing enzymes also present in liver microsomes comprise esterases, flavin monooxygenases (FMO), glucuronosyltransferases (UGT), glutathione S-transferase (GST) and methyl transferases (MT). The principal shortcoming associated to the use of a single enzymatic system or subcellular fractions arises in the necessity to direct the studies to a specific type of biotransformation, i.e. oxidation, glucuronidation, sulfation. . ., so that the overall metabolism cannot be simulated. Hence, depending on the biological system used, the set of metabolites will be different and, perhaps, they may not be the most relevant in vivo. Moreover, the incubation must be supplemented with the required cofactor(s) (Table 1). This drawback can be overcome carrying out the metabolic reactions with cells (e.g. cryopreserved hepatocytes and freshly isolated hepatocytes) or portions of an organ (e.g. freshly

Table 1 List of the most common reactions mediated by phase I and phase II enzymes and the required cofactors. Reaction

Enzyme

Cofactor/Reactive

Phase I Oxidation Oxidation Reduction Hydrolysis Phase II Glucuronidation Sulfation Methylation Acetylation Amino acid conjugation Glutathione conjugation

CYP450 FMO Alcohol and aldehyde dehydrogenase Xanthine dehydrogenase Reductive enzymes Esterase UGT SULT Methyltransferase Acetyltransferase Glutathione-S-transferase

NADPH NAD+ In general require NADPH (absence of O2 ) – UDP-Glucuronic acid (UDPGA) 3 -phosphoadenosine-5 -phosphosulfate (PAPS) S-adenosylmethionine (SAM) Acetyl-CoA Amino Acids Glutathione

106

J. Saurina, S. Sentellas / J. Chromatogr. B 1044 (2017) 103–111

isolated liver slices). These systems may provide conclusions more comparable to in vivo metabolism. In vivo assays rely on the administration of the drug to toxicological species such as rats and mice, and to humans in the case of clinical trials. Samples resulting from in vivo assays consist of biological fluids (e.g., plasma, bile, urine, etc.) which are analyzed to investigate the occurrence of metabolites as well as their evolution over time. Obviously, this is the most reliable system in metabolite profiling, however, the complexity of the samples makes more complicated their analysis and the interpretation of the data obtained. Hence, previous knowledge from in silico and in vitro tests is highly appreciated to facilitate a comprehensive assessment of drug metabolism. Although in vitro and in vivo assays are currently the methods of choice for metabolite profiling, other strategies are welcomed for the generation of drug metabolites in simpler or more massive ways. This is the case of the metabolite biosynthesis using different microbial strains [11–14]. The electrochemical generation of metabolites has been proposed as a new biomimetic strategy to complement conventional drug metabolism assays, especially for high-throughput screening of oxidative biotransformations [8,9]. Some drug processes that can be induced comprise N-dealkylation, S-oxidation, dehydrogenation, hydroxylation S-dealkylation and Noxidation reactions. [15–17].

3. Sample treatment Although this overview is not intended to cover exhaustively the sample treatment, some general considerations of this issue deserve our attention because this step is commonly needed prior to LC analysis. Metabolic samples from in vitro and in vivo assays often display a great complexity as they contain a wide range of components such as mineral elements and inorganic molecules, small organic compounds and biopolymers of molecular mass higher than 106 Da. Apart from this size variety, the diversity of components is also exceptional in terms of other physicochemical properties such as polarity, acidity constants, solubility in organic solvents, etc. As a result, the study of metabolites may result in a challenging task as multiple interferences and matrix effects may hinder the extraction of reliable conclusions. As commented below, the principal objectives of the sample treatment comprise the isolation and preconcentration of metabolites and, occasionally, their derivatization to enhance the sensitivity and selectivity of the study. Protein precipitation with organic solvents such as ACN is one of the most common preliminary treatments of the metabolic samples either from in vitro or in vivo assays. The supernatant phase containing metabolites can be analyzed directly by LC or subjected to additional cleanup procedures [18]. Liquid-liquid extraction (LLE), using organic solvents, is an excellent cleanup technique to recover metabolites free of small organic molecules and biopolymers. The extraction is facilitated by sonication or vortexing and the recovery of a clean extract by centrifugation or freezing. Solid phase extraction (SPE) purifies and pre-concentrates the metabolites according to the interactions stablished with the stationary phases of the extraction cartridges [19,20]. Reversed-phase C18 materials are widely used because of their ability to retain a broad range of organic molecules of intermediate and low polarity. Functionalized polymeric sorbents and ion exchange (cation or anion) resins have also been used for polar and ionic substances, respectively. The performance of these techniques will depend on the physicochemical characteristics of the obtained metabolites. In addition, occasional losses of metabolites because of adsorption processes and interactions with proteins cannot be underestimated as additional issues to be taken into account in these sample treatments.

Apart from the aforementioned methods, additional fraction collection from preparative or semi-preparative LC results in one of the most powerful approaches for the treatment of biological samples [19,20]. Time windows are defined to collect specific fractions containing the desired metabolite(s) which can be, afterwards, analyzed by structural elucidation techniques. Complementarily to physical cleanup and pre-concentration, chemical processes including analyte labelling and hydrolysis may be carried out. Derivatization reactions are used to enhance sensitivity of the detection, especially in the case of UV–vis absorption and fluorescence spectroscopies [21]. Hydrolysis reactions, either chemical or enzymatically, can be used to look into the occurrence of adducts from phase II conjugations [22–24]. For instance, Gupta et al. used the hydrolysis of glucuronide and sulfate conjugates to confirm the identity of hydrastine phase II metabolites [22]. In another example, He and coworkers identified some metenolone conjugates in urine by LC-QTOF after hydrolysis with glucuronidase [23].

4. Metabolite separation by liquid chromatography Metabolic extracts resulting from procedures given in Section 3 can be analyzed by separation techniques in order to resolve the metabolite profiles. The use of high performance separation will be fundamental to avoid problems of mutual co-elutions or interferences from other sample components. LC is one of the most important instrumental techniques for dealing with metabolite profiling in the field of drug discovery and development. Apart from LC, other separation techniques such as gas chromatography (GC) and capillary electrophoresis (CE) have been used occasionally in alternative studies. GC coupled to MS has been pointed out as another instrumental possibility due to its great resolution and elucidation performance [25–27]. Anyway, the field of application of GC–MS is less extended than that of the LC counterpart due to volatility issues of drugs and metabolites. This drawback has been sometimes solved via derivatization in which the analytes are transformed into less polar and more volatile derivatives. In this regard, some inappropriate functional groups such carboxylic acids and amines can be transformed into esters or amides, much more suitable for GC analysis. The application of CE with UV, FLD or, especially MS detection seemed to be highly promising due to its great separation performance [28]. The separation ability of CE is especially claimed in the case of charged drugs and metabolites in which the performance of LC is typically limited. The lower sensitivity exhibited by CE methods in comparison with LC counterparts may be incompatible with metabolite levels of some in vitro and in vivo assays. Fortunately, recent advances in on-line pre-concentration techniques relying on field-amplified sample stacking, sweeping and on line solid phase extraction processes have improved the limits of detection in 2 or 3 orders of magnitude, thus contributing to expand the applicability of CE. Anyway, despite the high expectations stirred up by GC and CE techniques, the continuous evolution of LC to get enhanced resolution, greater sensitivity and specificity, and high speed, makes LC (and especially the hyphenation LC–MS) the preferred platform for metabolite identification and determination. Hence, it should be remarked that our work will be mainly devoted to discuss the possibilities of LC for metabolite profiling. For decades, conventional high performance liquid chromatography (HPLC) resulted in the separation technique of choice for drug metabolism assays. In the last years, however, classical HPLC methods have become more residual in detriment of the newest and most powerful versions based on ultra-high performance liquid chromatography (UHPLC) [29] and core-shell column technology. Despite the higher cost of the instrumentation, UHPLC has quickly

J. Saurina, S. Sentellas / J. Chromatogr. B 1044 (2017) 103–111

(a) 45 0.04 5

Absorbance (mAU)

CIN M3

30 0.03 0

AU

M8

M9

15 0.01 5

* M1

*

0.000 0 0.00

5 5.00

10 10 .00

20 20.00

1515 .00

tes TimeMinu (min)

25 25.00

(b) 100 0.01

AU

Absorbance (mAU)

M3

*

CIN

* *

M4 M5

M1

M9

M2 M6

0.0000

2.00 2

4.00 4

6.00 6

two analytical columns of different nature, e.g. reversed-phase and HILIC, to achieve more exhaustive separations. In this technique, for instance, one dimension may be focused on a rougher separation of families of components while the second chromatography may be intended to a finer discrimination of related analytes. As a result, groups of poorly resolved compounds eluted along the first separation domain are injected into the second column system to reach a more specific resolution of analytes. For instance, Holland et al. used 2D-HPLC operating in heart-cutting mode to overcome the peak capacity limitations of the one-dimensional separation for the determination of neurotransmitters and their metabolites [35]. In another example, achiral and chiral separations were combined to separate albendazole metabolites from in vitro microsomal assays [36]. As a result, the optical isomers co-eluting in the first run could be separated in the chiral column. 4.1. Detectors in liquid chromatography

M8

55 0.00

107

8.00 8

M10

1010 .00

(min) Time t(min)

M12 M13 M15 12 .00 12

14 14.00

1616 .00

Fig. 3. Comparison of chromatograms obtained by (A) HPLC and (B) UHPLC. Sample obtained after incubation of cinitapride (20 mmol L−1 ) with human liver microsomes (mg protein mL−1 ) for 60 min. * Blank species.

gained popularity in our laboratories due to its great performance, thus allowing excellent separations in shorter analysis time compared to HPLC. As a result, some critical cases of structurally close metabolites that cannot be fully resolved by the traditional HPLC have been successively separated by UHPLC. Fig. 3 compares the chromatograms of a sample from an in vitro cinitapride assay by HPLC and UHPLC, showing that UHPLC provides a better resolution and analysis time [30]. As an alternative to the UHPLC, the core-shell technology also achieves great figures of merit working with conventional HPLC instruments. For instance, Hroch and coworkers developed an HPLC fluorescence method for the determination of boldine in plasma, bile and urine of rats using such a kind of columns. Complementary studies for the identification of its major metabolites relied on MS/MS [31]. In another study, Kloos et al. evaluated various types of core-shell columns differing in the stationary phases for fast urinary metabolic profiling [32]. A comprehensive comparison of characteristics of traditional, core-shell and UHPLC columns for metabolite profiling can be found in Ref. [33] by Dubbelman et al. In general, LC separations have been carried out using reversed phase mode, being C18 stationary phases the most widely used. Reversed phase columns are compatible with a wide range of drug polarities. Other stationary phases have been also introduced (for instance pentafluorophenyl, hydro-RP modifications) to stablish more efficient interactions with polar compounds thus leading to better separations of such a type of species. More recently, hydrophilic interaction liquid chromatography (HILIC) has been applied to conduct the separation of more polar compounds which, otherwise, would be hardly retained by the reversed phase counterpart [34]. As a more comprehensive and powerful approach, two-dimensional LC (2D-LC) has been introduced to enhance the separation capacity. 2D-LC may be suitable for the analysis of samples of high complexity containing species with a broad range of physicochemical characteristics. 2D-LC consists of the coupling of

Regarding detection in LC, ultraviolet (UV), fluorescence (FLD) and, especially, MS are commonly used for drug metabolism studies. Rather than conventional spectrophotometers, UV detectors for LC are often based on diode array spectrophotometers that allow to record full spectra throughout the chromatographic runs as well as chromatograms at one or several preselected wavelengths. Despite not being universal, UV detection is compatible with a wide variety of drugs, those displaying absorbing moieties in their molecules. In the case of FLD, the intrinsic values of sensitivities and LOD are excellent for detecting the low drug and metabolite concentrations occurring in the metabolic samples. Unfortunately, the number of drugs containing fluorogenic moieties is more limited. MS is currently one of the most useful techniques to deal with metabolic assays from both universality and sensitivity characteristics. Because of the great analytical possibilities of LC–MS, this hyphenation is discussed in detail below. New instrumental trends are focused on the development of LC NMR couplings for obtaining a more versatile and powerful platforms [37]. NMR detection is especially successful for resolving the structures of molecules because of its ability to deduce the exact positions of functional groups in the molecules which is, indeed, the recurrent problems of MS. In this regard, LC NMR may result in an excellent complement to LC–MS being 1 H and 13 C the principal nuclei to be considered. Some drawbacks to be solved concern sensitivity, rapid spectral acquisition throughout the chromatogram and use of deuterated mobile phases. However, as a significant advantage, the time-consuming sample preparation step can be simplified greatly as the separation process removes most of the matrix interferences. As example, Su et al. combine the capacity of LC-ESI–MSn and LC NMR for the structural elucidation of in vivo metabolites of isobavachalcone [38]. Additionally, the online hyphenation of LC, SPE and NMR, has been also demonstrated to be a suitable alternative for the structural elucidation process [39]. The inclusion of an on-line SPE step is a good strategy in order to increase the amount and purity of product entering the RMN cell leading to an improved sensitivity. LC with radiochemical detection is an excellent approach for the study of radioactivity-labelled drugs as well as for the generation of quantitative data on metabolites. Although several radioisotopes could be considered, 14 C or 3 H (tritium) are, by far, the most widely used since C and H are the basic elements of all organic drug molecules. LC-radiochemical methods are required for the unequivocal monitoring of drug related substances from in vitro or in vivo samples [40] and the determination of the mass balance of elimination and absorption [41–43]. The lack of structural information from radiochemical data, which is one of the main shortcomings, can be circumvented by complementary MS analysis.

108

J. Saurina, S. Sentellas / J. Chromatogr. B 1044 (2017) 103–111 1000

1000 Abundance

800

800

600 400

Abundance

200

600

0

0

10

20

30

t (min) M1

400

M2 M3

200

0

0

10

20

30

t (min) Fig. 4. LC-ICP–MS chromatograms (m/z 48) of rat bile sample obtained 4 h after intravenous ethacrynic acid (EA) administration (10 mg kg−1 ) to rats. Inside blank sample corresponding to the rat bile obtained before EA administration. M1, M2 and M3 correspond to EA-SG, EA-NAC and EA-Cys-Gly, respectively.

In the last years, LC-ICP–MS has appeared in the analytical scenario to deal with some drug metabolism studies. In this regard, drugs containing in their structures elements such as P, As, S, Se, Cl, Br, I, etc. are susceptible of being detected by ICP-MS [44–50]. Fig. 4 shows an example of LC-ICP-MS chromatograms (m/z 48) of rat bile sample obtained 4 h after intravenous ethacrynic acid (EA) administration (10 mg kg−1 ) to rats. In this case, S detection relies on monitoring SO+ ions. Unfortunately, the technique is not universal as H, C, N and O cannot be detected in this way so that most of the drugs and their corresponding metabolites are invisible to this technique. 5. Detection, structural elucidation and quantification of metabolites by LC–MS As indicated above, MS is one of the dominant detection techniques for the characterization of drug-related compounds in the pharmaceutical research, mainly using electrospray as the ion

source. The development of LC–MS over the last 50 years applied to drug metabolism has been reviewed elsewhere [51]. Although mass spectrometry can be used alone to record MS spectra, the coupling LC–MS provides additional advantages regarding the separation of components. This feature is especially relevant when various metabolites are formed simultaneously. Besides, LC–MS offers a great sensitivity compatible with those metabolites occurring at very low concentrations. Several LC–MS strategies can be followed for the identification of drug metabolites. As the starting point, the full scan MS analysis of control and test samples may provide overall information of ionizable metabolites. Besides, the functionalization of the molecule by comparison with the parent drug can also be inferred. This is because the formation of a given metabolite commonly involves a mass shift associated to the gain or loss of one or several atoms with respect to the parent drug (Table 2). Moreover, computer-based algorithms, based on mass shifts, have been implemented in most of MS softwares, thus making easier and faster the identification of the drug candidate biotransformation. However, the detection of metabolites by full scan may be especially hard when dealing with complex biological samples such as urine or plasma because a multitude of ions from non-drug-related components can provoke interferences with the metabolite MS signals. The information extracted from the full scan experiments is often sufficient to deduce the transformations that the parent compound has underwent although it is rarely sufficient to ascertain the exact position of such a change. Richer structural information can be elucidated from MS/MS experiments in which a given ion is captured, fragmented and the resulting daughter ions are detected in the MS analyzer. The interpretation of MS/MS spectra of metabolite and parent compound may contribute to infer the structure of the metabolite or, at least, to reduce the number of possibilities in the functionalization position. This is the case, for instance, of metabolites of aclidinium bromide in which we observed a metabolite (M1) with a shift mass of 16 Da with respect the metabolite M2 (unequivocally identified). In this case, the unique possibility for this mass shift was a hydroxylation. Comparing the MS/MS spectra of both parent compound (M2) and M1 the position of the hydroxylation can be attributed to a phenyl ring. To reach to this conclusion

Table 2 Mass shifts associated to the gain or loss of one or several atoms with respect to the parent drug. Mass shift Phase I +15.9949

Metabolic reaction

Example

Hydroxylation, epoxidations, oxidations to N-oxides, S-oxides and sulfones, oxidation of aldehydes to carboxylic acids

R-CH2 → R-CHOH R-HC CH-R → R-CHOCH-R R-S-R → R-SO-R R-S-R → R-SO2 -R R-CH2 -OH → R-COOH R-O-CH3 → R-OH R-NH-CH3 → R-NH2 R-S-CH3 → R-SH R-CH2 -OH → R-COH

+31.9898 +13.9792 −14.0156

Addition of two oxygen Oxidation of alcohol to carboxylic acid Oxidative demethylation (N-, O-, S-dealkylation)

−2.0156 −29.9741 +18.0105

Formation of an unsaturated bond Oxidation of alcohols to aldehides N reduction Epoxide hydratation

Phase II +176.0320

Glucuronidation

+305.0681/307.0837 +161.0147/163.0303 +79.9568 +14.0156 +42.0105 +57.0214 +71.0371 +114.0793

Addition of glutathione (GSH) - Glutathione - Mercapturic acid Sulfation Methylation Acetylation Amino acid conjugation - Glycine - Alanine - Ornithine

R-NO2 → R-NH2 R-CH(O)CH-R → R-CH(OH)-CH(OH)-R R-OH → R-O-Glu R-NH2 → R-NH-Glu R-CH CH2 → R-CH2 -CH2 -SG R-CH CH2 → R-CH2 -CH2 -S-mercapturic acid R-OH → R-OSO3 H R-OH → R-OCH3 R-NH2 → R-NH-CO-CH3 R-COOH → R-CO-Gly R-COOH → R-CO-Ala R-COOH → R-CO-Orn

J. Saurina, S. Sentellas / J. Chromatogr. B 1044 (2017) 103–111 Table 3 Typical fragments used in metabolite characterization by neutral loss (NL) and product ion (PI) scan modes. Metabolite

Mode

Scan

Glucuronide Phenolic sulfates Aliphatic sulfates Aliphatic GSH adducts Aryl GSH adducts GSH adducts

+/− + – + + –

NL 176 (-C6 H8 O6 ) NL 80 (-SO3 ) PI 97 (HSO4 ) NL 129 (-C5 H7 NO3 ) NL 275 (-C10 H17 N3 O6 ) PI 272 (C10 H14 N3 O6 )

MS/MS fragments obtained for M1 and M2 were studied. Some fragments were identical for both compounds while others had a difference in mass of 16 Da. The identification of the fragments was then crucial to ascertain which moiety of the molecule was modified. However, with this approximation the exact position of hydroxylation was not ascertained and further studies based on NMR were necessary [52]. Similarly, MSn data obtained from ion trap analyzers is very useful to clarify the structure of metabolites formed. Anyway, in the case of multiple functional isomers some ambiguities in the location of the functional groups may persist. Triple quadrupole mass spectrometer-based neutral loss (NL) and precursor ion (PI) scans are common approaches to determine unknown metabolites. The former, relies on monitoring those compounds that show a specific loss of a characteristic neutral fragment. For instance, neutral losses of 176 or 80 are characteristic of glucuronide conjugates or sulfates, respectively. Likewise, in precursor ion scan mode, all the compounds providing a common product ion are detected. In Table 3, a list of the most typical fragments to be used in metabolite elucidation by NL and PI is given. The main drawback of these scan modes is their rather low sensitivity. This issue can be overcome using the multiple reaction monitoring (MRM) mode which provides the best sensitivities for MS-based determinations. However, this acquisition mode relies on the prior knowledge of the metabolites to be searched to define the monitoring conditions. Otherwise, unknown or non-expected metabolites may be overlooked since conditions for their detection will be not specified in the measurement parameters. The evolution of MS over the last decade has led to high resolution mass spectrometry (HRMS) using time-of-flight (TOF) or Orbitrap mass analyzers. HRMS instruments have a tremendous impact in the field of metabolite profiling [53]. Nowadays, HRMS is not only used for an accurate mass determination of the molecular ion and their fragments but also for opening up new opportunities in the detection and identification of metabolites based on specific MS modes. With the MSE mode all components are fragmented in a single run. The global dataset can be used for metabolite characterization using different post-acquisition filters such as neutral loss filtering (NLF) and precursor ion filtering (PIF). Another softwarebased data processing method is the so-called mass defect filter (MDF) relying on the difference between the exact mass and the nominal mass of the compound. MDF has been gained popularity in the last years in the field of drug metabolism research due to the fact that the mass defect of metabolite ions commonly fall within 50 mDa relative to that of the parent compound. Post-acquisition data mining strategies, especially the application of MDF, were demonstrated for in vivo metabolite profiling using compounds AZD6280 and AZD12488024 as model drugs [54]. Other examples dealing with combinations of different post-acquisition methods can be found in references [55–58]. In addition to the above commented strategies, the evolution of MS softwares has allowed the samples to be monitored by information-dependent acquisition (IDA) mode. IDA combines two different scan modes in a unique run using data obtained with a given acquisition mode (e.g., single scan) to define parameters of a second acquisition mode. Hence, when a given peak deserves

109

special attention according to criteria that the user has prestablished, the peak is examined more deeply with the second MS mode, so additional data is given. The suitability of this strategy has been demonstrated in the determination of microsomal metabolism of amitriptyline and praziquantel by combining predictive multiple reaction monitoring – information-dependent acquisition – enhanced product ion (MRM-IDA-EPI) scanning [59,60]. The use of isotopic clusters to identify the metabolites is relatively efficient for those molecules containing atoms with a characteristic ion patterns, such as Cl (35 Cl and 37 Cl isotopes with abundancies of 67 and 33%, respectively). However, these atoms are not present in a vast majority of drugs so, the direct use of isotopic cluster monitoring may be limited. In order to expand the applicability of this approach, the incorporation of stable isotopes, such as deuterium and 13 C, into the drug molecules opens up new opportunities. Under the so-called isotopic dilution method, the dilution of the original drug with the stable isotope labelled compound provides characteristic spectra with a twin ion pair pattern (usually separated by >2 Da) for the parent drug and all related substances. When these drug mixtures are used for in vivo administration or in vitro incubation, the tracking of this characteristic pattern facilitates the identification of metabolites in complex matrices [61,62]. Apart from the modification of the parent drug with stable isotopes, they can also be incorporated from labelling molecules via drug derivatization. This is the case, for instance, of isotopic drug labelling through the use of GSH-labelled as a reagent [63,64]. As a different strategy, deuterated solvents can also be exploited for the identification of metabolites based on H/D exchange. Hence, the number of labile hydrogen atoms in the analyte molecule can be investigated in this way. The procedure consists of the treatment of the sample with D2 O as the solvent. Labile hydrogens corresponding to alcohols, phenols, organic acids, amines or thiols are then exchanged by deuterium, producing a shift of +1 Da per H substituted. Besides, this strategy has been successfully used for the differentiation between hydroxylated and N-oxide metabolites [30,65]. Other emerging trends such as imaging mass spectrometry or ambient ionization techniques have also been applied for metabolite profiling [66]. In the last years, the combination of ion mobility spectrometry (IMS) with MS detection (IMS-MS) has gained popularity in various analytical areas. IMS separates ionic species in an electric field based on their different mobilities, which depend on the m/z values and the collision cross-section (CCS) of the ions. Since CCS is a physical feature that reflects the shape of the compounds, the comparison of experimental or calculated CCS values of target analytes with standard compounds can be used to study the structure of ions. In this direction, Shimizu and coworkers demonstrated the suitability of IMS-MS for the determination of the site of glucuronidation or aromatic hydroxylation of new drug candidates [67,68].

6. Chemometrics for the treatment of LC data The tremendous amounts of data obtained from the metabolic research, especially in the case of LC–MS, result in an exceptional source of information deserving a thorough analysis. In this section we describe some chemometric applications for a more comprehensive evaluation of metabolic profiling. Studies are typically conducted under the so-called metabolomic approach in which two sets of individuals/samples differing in the pharmacological treatment administered are compared on the basis of the metabolic profiles. Chen and coworkers published an interesting overview on the application of metabolomics to the study of drug metabolism [69].

110

J. Saurina, S. Sentellas / J. Chromatogr. B 1044 (2017) 103–111

Data resulting from in vitro and in vivo assays can be analyzed chemometrically to try to recover the underlying features of metabolites [70]. Among others, principal component analysis (PCA) is the most widely used method for exploratory studies to investigate the relationships among chemical data and individual/sample features. Complementarily, other methods such as partial least squares (PLS), partial least squares − discriminant analysis (PLS-DA) and soft-independent modeling of class analogy (SIMCA) can be used for the classification of samples into predefined groups such as treated and untreated individuals. In the example by Schneider et al., PCA was applied to study the in vitro metabolism of nefazodone [71]. Samples from human liver microsome assays incubated in the presence and without the drug were analyzed by LC–MS using a high throughput screening method. Data treatment with PCA revealed the segregation among treated and non-treated samples. The plot of loadings allowed the ions corresponding to drug and metabolites to be encountered. Further MS/MS analysis of targeted ions permitted the identification and elucidation of the molecular structures of nefazodone and its metabolites. Analogous PCA-based approaches for the comparison of control and drug-treated groups were used for the assessment of the in vivo metabolism of atazanavir in mice [72] and cocaine in mice and rats [73]. 7. Conclusions/further trends The preliminary identification and evaluation of metabolites are fundamental in the research of new drug candidates to evaluate aspects dealing with activity and toxicity. Mass spectrometry in combination with liquid chromatography (LC–MS) is currently the most popular technique for drug metabolism studies. Among other advantages, LC–MS provides high separation capacity, high sensitivity compatible with the detection of species occurring at very low levels, and structural information from the interpretation of MS and MS/MS spectra. Regarding other detectors, LC NMR is envisaged as the complementary technique to LC–MS, especially for structural elucidation of unknown metabolites. The success of the first applications of chemometrics and metabolomic-based approaches to the study of drug metabolism will entail the development of these disciplines for more comprehensive characterizations of metabolic systems. Acknowledgements This work was partially financed by the Generalitat de Catalunya and the Spanish National Plan of Research with projects NASSOS 2014SGR377 and Superfactory CTQ2014-56324. References [1] F.J. Morales-Gutierrez, J. Barbosa, D. Barron, Metabolic study of enrofloxacin and metabolic profile modifications in broiler chicken tissues after drug administration, Food Chem. 172 (2015) 30–39. [2] S.M. Attia, Deleterious effects of reactive metabolites, Oxid. Med. Cell. Longev. 3 (2010) 238–253. [3] H. Takakusa, H. Masumoto, H. Yukinaga, C. Makino, S. Nakayama, O. Okazaki, K. Sudo, Covalent binding and tissue distribution/retention assessment of drugs associated with idiosyncratic drug toxicity, Drug Metab. Dispos. 36 (2008) 1770–1779. [4] J. Svenson, V. Vergote, R. Karstad, C. Burvenich, J.S. Svendsen, B. De Spiegeleer, Metabolic fate of lactoferricin-based antimicrobial peptides: effect of truncation and incorporation of amino acid analogs on the in vitro metabolic stability, J. Pharmacol. Exp. Ther. 332 (2010) 1032–1039. [5] Guidance for industry, in: Safety Testing of Drug Metabolites, U.S. Department of Health and Human Services. Food and Drug Administration. Center for Drug Evaluation and Research (CDER), 2008 http://www.fda.gov/OHRMS/ DOCKETS/98fr/FDA-2008-D-0065-GDL.pdf. [6] D. Zhanga, G. Luob, X. Dingc, C. Lud, Preclinical experimental models of drug metabolism and disposition in drug discovery and development, Acta Pharm. Sin. 2 (2012) 549–561.

[7] L. Jia, X. Liu, The conduct of drug metabolism studies considered good practice (II): in vitro experiments, Curr. Drug Metab. 8 (2007) 822–829. [8] W. Lohmann, U. Karst, Biomimetic modeling of oxidative drug metabolism, Anal. Bioanal. Chem. 391 (2008) 79–96. [9] A. Baumann, U. Karst, Online electrochemistry/mass spectrometry in drug metabolism studies: principles and applications, Expert Opin. Drug Met. 6 (2010) 715–731. [10] C.H. Andrade, D.C. Silva, R.C. Braga, In silico prediction of drug metabolism by P450, Curr. Drug Metab. 15 (2014) 514–525. [11] A. Long, Drug metabolism in silico – the knowledge-based expert system approach. Historical perspectives and current strategies, Drug Discov. Today 10 (2013) e147–153. [12] D.L. Zhang, H.Y. Zhang, N. Aranibar, R. Hanson, Y. Huang, P.T. Cheng, S. Wu, S. Bonacorsi, M.S. Zhu, A. Swaminathan, W.G. Humphreys, Structural elucidation of human oxidative metabolites of muraglitazar: use of microbial bioreactors in the biosynthesis of metabolite standards, Drug Metab. Dispos. 34 (2006) 267–280. [13] M. Geier, A. Braun, A. Emmerstorfer, H. Pichler, A. Glieder, Production of human cytochrome P450 2D6 drug metabolites with recombinant microbes – a comparative study, Biotechnol. J. 7 (2012) 1346–1358. [14] A. Osorio-Lozada, S. Surapaneni, G.L. Skiles, R. Subramanian, Biosynthesis of drug metabolites using microbes in hollow fiber cartridge reactors: case study of diclofenac metabolism by Actinoplanes species, Drug Metab. Dispos. 36 (2008) 234–240. [15] T. Johansson, L. Weidolf, U. Jurva, Mimicry of phase I drug metabolism–novel methods for metabolite characterization and synthesis, Rapid Commun. Mass Spectrom. 21 (2007) 2323–2331. [16] A. Baumann, A. Faust, M.P. Law, M.T. Kuhlmann, K. Kopka, M. Schafers, U. Karst, Metabolite identification of a radiotracer by electrochemistry coupled to liquid chromatography with mass spectrometric and radioactivity detection, Anal. Chem. 83 (2011) 5415–5421. [17] S. Jahn, A. Baumann, J. Roscher, K. Hense, R. Zazzeroni, U. Karst, Investigation of the biotransformation pathway of verapamil using electrochemistry/liquid chromatography/mass spectrometry. A comparative study with liver microsomes, J. Chromatrogr. A 1218 (2011) 9210–9220. [18] L. Taevernier, E. Wynendaele, M. D’Hondt, B. De Spiegeleer, Analytical quality-by-design approach for sample treatment of BSA-containing solutions, J. Pharm. Anal. 5 (2015) 27–32. [19] B. Alvarez-Sanchez, F. Priego-Capote, M.D. Luque de Castro, Metabolomics analysis I. Selection of biological samples and practical aspects preceding sample preparation, Trends Anal. Chem. 29 (2010) 111–119. [20] B. Alvarez-Sanchez, F. Priego-Capote, M.D. Luque de Castro, Metabolomics analysis II. Preparation of biological samples prior to detection, Trends Anal. Chem. 29 (2010) 120–127. [21] B.L. Qi, P. Liu, Q.Y. Wang, W.J. Cai, B.F. Yuan, Y.Q. Feng, Derivatization for liquid chromatography-mass spectrometry, Trends Anal. Chem. 59 (2014) 121–132. [22] P.K. Gupta, G. Barone, B.J. Gurley, E.K. Fifer, H.P. Hendrickson, Hydrastine pharmacokinetics and metabolism after a single oral dose of goldenseal (Hydrastis canadensis) to humans, Drug Metab. Dispos. 43 (2015) 534–552. [23] G.Y. He, S. Yang, J.H. Lu, Y.X. Xu, New long term metabolite in human urine for metenolone misuse by liquid chromatography quadrupole time-of-flight mass spectrometry, Steroids 105 (2016) 1–11. [24] A. Knoop, O. Krug, M. Vincenti, W. Schanzer, M. Thevis, In vitro metabolism studies on the selective androgen receptor modulator LG121071 and its implementation into human doping controls using liquid chromatography-mass spectrometry, Eur. J. Mass Spectrom. 21 (2015) 27–36. [25] P. Kavanagh, A. Grigoryev, A. Melnik, S. Savchuk, A. Simonov, V. Rozhanets, Detection and tentative identification of urinary phase I metabolites of phenylacetylindole cannabimimetics JWH-203 and JWH-251, by GC-MS and LC-MS/MS, J. Chromatogr. B 934 (2013) 102–108. [26] X. Matabosch, O.J. Pozo, C. Perez-Mana, M. Farre, J. Marcos, J. Segura, R. Ventura, Identification of budesonide metabolites in human urine after oral administration, Anal. Bioanal. Chem. 404 (2012) 325–340. [27] Q. Li, G.F. Zhang, Identification of n-hydroxy acid metabolites in electron impact ionization mass spectrometry, Rapid Commun. Mass Spectrom. 26 (2012) 1355–1362. [28] A. Baldacci, F. Prost, W. Thormann, Identification of diphenhydramine metabolites in human urine by capillary electrophoresis-ion trap-mass spectrometry, Electrophoresis 25 (2004) 1607–1614. [29] D. Spaggiari, L. Geiser, S. Rudaz, Coupling ultra-high-pressure liquid chromatography with mass spectrometry for in-vitro drug-metabolism studies, Trends Anal. Chem. 63 (2014) 129–139. [30] H. Marquez, J. Alberti, M. Salva, J. Saurina, S. Sentellas, Development of a UHPLC method for the assessment of the metabolic profile of cinitapride, J. Sep. Sci. 24 (2011) 3502–3508. [31] M. Hroch, S. Micuda, J. Cermanova, J. Chladek, P. Tomsik, Development of an HPLC fluorescence method for determination of boldine in plasma, bile and urine of rats and identification of its major metabolites by LC-MS/MS, J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 936 (2013) 48–56. [32] D.P. Kloos, H. Lingeman, W.M.A. Niessen, A.M. Deelder, M. Giera, O.A. Mayboroda, Evaluation of different column chemistries for fast urinary metabolic profiling, J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 927 (2013) 90–96. [33] A.C. Dubbelman, F. Cuyckens, L. Dillen, G. Gross, T. Hankemeier, R.J. Vreeken, Systematic evaluation of commercially available ultra-high performance

J. Saurina, S. Sentellas / J. Chromatogr. B 1044 (2017) 103–111

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

[51] [52]

liquid chromatography columns for drug metabolite profiling: optimization of chromatographic peak capacity, J. Chromatogr. A 1374 (2014) 122–133. A. Ekdahl, A. Aurell-Holmberg, N. Castagnoli, Identification of the metabolites of lesogaberan using linear trap quadrupole orbitrap mass spectrometry and hydrophilic interaction liquid chromatography, Xenobiotica 43 (2013) 461–467. B.J. Holland, X.A. Conlan, P.G. Stevenson, S. Tye, A. Reker, N.W. Barnett, J.L. Adcock, P.S. Francis, Determination of neurotransmitters and their metabolites using one- and two-dimensional liquidchromatography with acidic potassium permanganate chemiluminescence detection, Anal. Bioanal. Chem. 406 (2014) 5669–5676. K.R.A. Belaz, E.R. Pereira-Filho, R.V. Oliveira, Development of achiral and chiral 2D HPLC methods for analysis of albendazole metabolites in microsomal fractions using multivariate analysis for the in vitro metabolism, J. Chromatogr. B 932 (2013) 26–33. Z. Yang, Online hyphenated liquid chromatography-nuclear magnetic resonance spectroscopy-mass spectrometryfor drug metabolite and nature product analysis, J. Pharm. Biomed. Anal. 40 (2006) 516–527. S. Su, Y.N. Wang, L. Bai, B.B. Xia, X.R. Li, Y. Tang, P.X. Xu, M. Xue, Structural elucidation of in vivo metabolites of isobavachalcone in rat by LC-ESI-MSn and LC-NMR, J. Pharm. Biomed. Anal. 104 (2015) 38–46. F. Gillotin, P. Chiap, M. Frederich, J.C. Van Heugen, P. Francotte, P. Lebrun, B. Pirotte, P. de Tullio, Coupling of liquid chromatography/tandem mass spectrometry and liquid chromatography/solid-phase extraction/NMR techniques for the structural identification of metabolites following in vitro biotransformation of SUR1-selective ATP-sensitive potassium channel openers, Drug Metab. Dispos. 38 (2010) 232–240. M. Kiffe, D.G. Schmid, G.J.M. Bruin, Radioactivity detectors for high-performance liquid chromatography in drug metabolism studies, J. Liq. Chromatogr. Relat. Technol. 31 (2008) 1593–1619. G.D. Bowers, A. Culp, M.J. Reese, G. Tabolt, L. Moss, S. Piscitelli, P. Huynh, D. Wagner, S.L. Ford, E.P. Gould, R. Pan, Y. Lou, D.A. Margolis, W.R. Spreen, Disposition and metabolism of cabotegravir: a comparison of biotransformation and excretion between different species and routes of administration in humans, Xenobiotica 46 (2016) 147–162. L.Z. Chen, A. Jungnik, Y.P. Mao, E. Philip, D. Sharp, A. Unseld, L. Seman, H.J. Woerle, S. Macha, Biotransformation and mass balance of the SGLT2 inhibitor empagliflozin in healthy volunteers, Xenobiotica 45 (2015) 520–529. M. Reyes, M. Hoch, P. Brossard, W. Wagner-Redeker, T. Miraval, J. Dingemanse, Mass balance, pharmacokinetics and metabolism of the selective S1P(1) receptor modulator ponesimod in humans, Xenobiotica 45 (2015) 139–149. C. Duckett, M. McCullagh, C. Smith, I.D. Wilson, The metabolism of 4-bromoaniline in the bile-cannulated rat: application of ICPMS (Br-79/81) HPLC-ICPMS & HPLC-oaTOFMS, Xenobiotica 45 (2015) 672–680. S. Stice, G.L. Liu, S. Matulis, L.H. Boise, Y. Cai, Determination of multiple human arsenic metabolites employing high performance liquid chromatography inductively coupled plasma mass spectrometry, J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 1009 (2016) 55–65. J.S.B. de Vlieger, M.J.N. Giezen, D. Falck, C. Tump, F. van Heuveln, M. Giera, J. Kool, H. Lingeman, J. Wieling, M. Honing, H. Irth, W.M.A. Niessen, High temperature liquid chromatography hyphenated with ESI-MS and ICP-MS detection for the structural characterization and quantification of halogen containing drug metabolites, Anal. Chim. Acta 698 (2011) 69–76. C. Losada, J.J. Alberti, J. Saurina, S. Sentellas, Determination of S-containing drug metabolites from in vitro and in vivo metabolism studies by using LC-ICP/MS, Anal. Bioanal. Chem. 404 (2012) 539–551. M. Wind, W.D. Lehmann, Element and molecular mass spectrometry – an emerging analytical dream team in the life sciences, J. Anal. At. Spectrom. 19 (2004) 20–25. B. Gammelgaard, B.P. Jensen, Application of inductively coupled plasma mass spectrometry in drug metabolism studies, J. Anal. At. Spectrom. 22 (2007) 235–249. K. De Wolf, L. Balcaen, E. Van De Walle, F. Cuyckens, F. Vanhaecke, ICP-MS A comparison between HPLC-dynamic reaction cell-ICP-MS and HPLC-sector field-ICP-MS for the detection of glutathione-trapped reactive drug metabolites using clozapine as a model compound, J. Anal. At. Spectrom. 25 (2010) 419–425. B. Wen, M. Zhu, Applications of mass spectrometry in drug metabolism: 50 years of progress, Drug Metab. Rev. 47 (2015) 71–87. J.J. Alberti, S. Sentellas, M. Salva, In vitro liver metabolism of aclidinium bromide in preclinical animal species and humans: identification of the

[53]

[54]

[55]

[56]

[57]

[58]

[59]

[60]

[61]

[62]

[63]

[64]

[65]

[66]

[67]

[68]

[69] [70] [71]

[72] [73]

111

human enzymes involved in its oxidative metabolism, Biochem. Pharmacol. 81 (2011) 761–776. M. Zhu, H. Zhang, W.G. Humphreys, Drug metabolite profiling and identification by high-resolution mass spectrometry, J. Biol. Chem. 286 (2011) 25419–25425. J. Guo, M. Zhang, C.S. Elmore, K. Vishwanathan, An integrated strategy for in vivo metabolite profiling using high-resolution mass spectrometry based data processing techniques, Anal. Chim. Acta 780 (2013) 55–64. Y. Liang, G. Wang, L. Xie, L.S. Sheng, Recent development in liquid chromatography/mass spectrometry and emerging technologies for metabolite identification, Curr. Drug Metab. 12 (2011) 329–344. J.E. Barbara, J.M. Castro-Perez, High-resolution chromatography/time-of-flight MS(E) with in silico data mining is an information-rich approach to reactive metabolite screening, Rap. Commun. Mass Spectrom. 25 (2011) 3029–3040. Y. Jin, C.S. Wu, J.L. Zhang, Y.F. Li, A new strategy for the discovery of epimedium metabolites using high-performance liquid chromatography with high resolution mass spectrometry, Anal. Chim. Acta 768 (2013) 111–117. A. Wohlfarth, A.S. Gandhi, S.K. Pang, M.S. Zhu, K.B. Scheidweiler, M.A. Huestis, Metabolism of synthetic cannabinoids PB-22 and its 5-fluoro analog 5F-PB-22, by human hepatocyte incubation and high-resolution mass spectrometry, Anal. Bioanal. Chem. 406 (2014) 1763–1780. J.Y. Lee, S.Y. Lee, K. Lee, S.J. Oh, S.K. Kim, Determination of species-difference in microsomal metabolism of amitriptyline using a predictive MRM-IDA-EPI method, Chem.-Biol. Interact. 229 (2015) 109–118. J. Huang, S.P.R. Bathena, Y. Alnouti, Metabolite profiling of praziquantel and its analogs during the analysis of in vitro metabolic stability using information-dependent acquisition on a hybrid triple quadrupole linear ion trap mass spectrometer, Drug Metab. Pharmacokinet. 25 (2010) 487–499. D. Gao, X.W. Chen, X.M. Yang, Q. Wu, F. Jin, H.L. Wen, Y.Y. Jiang, H.X. Liu, Stable isotope labeling strategy for curcumin metabolite study in human liver microsomes by liquid chromatography-tandem mass spectrometry, J. Am. Soc. Mass Spectrom. 26 (2015) 686–694. M.G. Leeming, A.P. Isaac, B.J. Pope, N. Cranswick, C.E. Wright, J. Ziogas, R.A.J. O’Hair, W.A. Donald, High-resolution twin-Ion metabolite extraction (HiTIME) mass spectrometry: nontargeted detection of unknown drug metabolites by isotope labeling, liquid chromatography mass spectrometry, and automated high-performance computing, Anal. Chem. 87 (2015) 4104–4109. I. Mezine, C. Bode, B. Raughley, S. Bhoopathy, K.J. Roberts, A.J. Owen, I.J. Hidalgo, Application of exogenous mixture of glutathione and stable isotope labeled glutathione for trapping reactive metabolites in cryopreserved human hepatocytes. Detection of the glutathione conjugates using high resolution accurate mass spectrometry, Chem.-Biol. Interact. 204 (2013) 173–184. T. Yamaoka, Y. Kitamura, Characterization of a highly sensitive and selective novel trapping reagent stable isotope labeled glutathione ethyl ester, for the detection of reactive metabolites, J. Pharm. Toxicol. Methods 76 (2015) 83–95. R.P. Shah, A. Garg, S.P. Putlur, S. Wagh, V. Kumar, V. Rao, S. Singh, S. Mandlekar, S. Desikan, Practical and economical implementation of online H/D exchange in LC-MS, Anal. Chem. 85 (2013) 10904–10912. W.J. Lu, Y.Z. Xu, Y.L. Zhao, X.B. Cen, Emerging technologies recent developments, and novel applications for drug metabolite identification, Curr. Drug Metab. 15 (2014) 865–874. A. Shimizu, T. Ohe, M. Chiba, A novel method for the determination of the of glucuronidation by ion mobility spectrometry-mass spectrometry, Drug. Metab. Spectrom. 40 (2012) 1456–1459. A. Shimizu, M. Chiba, Ion mobility spectrometry-mass spectrometry analysis for the site of aromatic hydroxylation, Drug Metab. Dispos. 41 (2013) 1295–1299. C. Chen, F.J. Gonzalez, J.R. Idle, LC-MS-based metabolomics in drug metabolism, Drug Metab. Rev. 39 (2007) 581–597. J. Trygg, E. Holmes, T. Lundstedt, Chemometrics in metabonomics, J. Proteome Res. 6 (2007) 469–479. R.P. Schneider, H. Zhang, L. Mu, A.S. Kalgutkar, R. Bonner, Utility of multivariate analysis in support of in vitro metabolite identification studies: retrospective analysis using the antidepressant drug nefazodone, Xenobiotica 40 (2010) 262–274. F. Li, J. Lu, L.Y. Wang, X.C. Ma, CYP3A-mediated generation of aldehyde and hydrazine in atazanavir metabolism, Drug Metab. Dispos. 39 (2011) 394–401. D. Yao, X.L. Shi, L. Wang, B.A. Gosnell, C. Chen, Characterization of differential cocaine metabolism in mouse and rat through metabolomics-guided metabolite profiling, Drug Metab. Dispos. 41 (2013) 79–88.